Study on Reliability Forecast of Gantry of Substation Based on Neural Network

摘要

This paper researches on the reliability forecast on two-span steel structural gantries of substation with unilateral support, bilateral supports and non-lateral support based on neural network. For the limited bearing capacity of the structure, the structural element with and without damage are taken into consided. The action on the structure Three kinds of action are taken into account, oblique load, horizontal load and vertical load. Then characteristics of second-order effect and influencing factors of gantry of substation for each case are computed. Afterwards, reliability forecast model of steel structural gantry of substation with high precision is established by artificial neural network method. The model can forecast elastic, elastic-plastic ultimate load-bearing capacity and corresponding deformation of gantry of substation with different damage degree effectively, so as to monitor and analyze safety of actual gantry of substation.